Laboratory of Genetics, Division of Basic Sciences, National Cancer Institute, National Institutes of Health, 37 Center Drive, Bethesda, MD 20892-4255, USA

Correspondence to: S Janz, Fax: 301 402 1031

DNA sequence analysis of PCR amplified Igh/c-myc junction fragments of T(12;15) chromosome translocations and immunohistochemical determination of immunoglobulin isotype production were employed to study the clonal diversification of neoplastic translocated plasma cells that resided in peritoneal inflammatory granulomas of BALB/c mice harboring primary plasmacytomas. The diversity of plasma cells was found to take two major forms when the fine structure of the T(12;15) translocation was used as the clonotypic marker. First, mosaics of clones containing translocations that were apparently unrelated to each other were detected in nine out of 17 (53%) mice. Second, subclones derived from common T(12;15)⁺ progenitors by either secondary deletions in translocation breakpoint regions or aberrant isotype switching near translocation breaksites were found in five of 17 (29.5%) mice. When Ig expression was utilized as the clonotypic marker, clonal mosaics were shown to occur in all mice. This was demonstrated by the finding that the prevalent IgA- or IgG-producing plasmacytoma clone was invariably accompanied by smaller clones of IgG- or IgA-expressing neoplastic plasma cells, respectively. These results provided new insights into the clonal diversification at the terminal stage of plasmacytomagenesis. In addition, they suggested that BALB/c plasmacytomas may be uniquely useful for studying clonal diversity during B cell oncogenesis, since clonal evolution can be evaluated in a pool of tumor and tumor precursor cells that is clearly defined by the T(12;15) chromosomal translocation and the production of monoclonal immunoglobulin. Leukemia (2000) 14, 909-921.

illegitimate genetic recombination; Igh/c-myc junctions; clonal evolution; abnormal isotype switching

Introduction

The clonal diversity of human plasma cell neoplasias at early preclinical stages of tumor development is largely undefined. Multiple myeloma, for example, is usually diagnosed as a monoclonal disease; yet, in the great majority of cases it remains unknown whether additional, subordinate clones of malignant plasma cells accompany the disease-causing clone at the time of diagnosis. Furthermore, in most cases it is not possible to decide whether the multiple myeloma originated as a monoclonal proliferation of a single aberrant precursor clone that developed in the absence of competing clones, or - in case multiple precursors did compete during myelomagenesis - when the myeloma clone emerged as the dominant cancer clone. Mouse models of plasma cell malignancies offer experimental systems in which problems of the sort mentioned above can be addressed in ways impossible to pursue in humans. Based on this premise, we decided to study clonal diversity during plasmacytomagenesis in the mouse, using peritoneal plasmacytomas that can be induced in genetically susceptible BALB/c mice as the experimental system. One advantage of this mouse model is the possibility to co-detect multiple aberrant clones during tumor development and to unambiguously identify tumor cells (plasmacytomas) and precursor B cells by virtue of the T(12;15) chromosome translocation. The T(12;15) is the hallmark mutation of BALB/c plasmacytomas, which results in the constitutive expression of the proto-oncogene c-myc.¹

In previous work we showed that the T(12;15) translocation comprises a very early and, most likely, initiating oncogenic event in the development of inflammation-induced peritoneal plasmacytomas.² Furthermore, the analysis of peritoneal inflammatory granulomas, the tissue site where plasmacytomas develop,^3,4 revealed that numerous T(12;15)⁺ clones were commonly present during tumor latency long before frank malignant growth was observed.^5,6,7 These results suggested that translocated cells residing in inflammatory granulomata constitute a remarkably large and diverse pool of tumor precursor cells from which plasmacytomas could evolve. However, our observations on the co-existence of multiple T(12;15)⁺ clones in the initial stages of oncogenesis did not address the question of whether a similarly large number of translocated clones was present in the late stages of plasmacytomagenesis. Here we attempt to answer this question by focusing on the end-point of oncogenesis in mice that completed tumor development and harbored incipient plasmacytomas. We demonstrate that multiple clones of translocated cells are frequently present in BALB/c mice that bear early, intermediate, or even advanced plasmacytomas. We further show that mouse plasmacytomagenesis seems to be characterized by the vigorous generation of subclones derived from common T(12;15)⁺ progenitors. The generation of subclones takes several forms, but one peculiar form of clonal diversification, detected in as many as five out of 17 (29.5%) plasmacytomas, appeared to be particularly noteworthy: the 'deletional remodeling' of translocation breakpoint regions by aberrant isotype switching.

Supported by these new data, we propose to consider mouse plasmacytoma development as a model of malignant B cell transformation that is characterized by (1) the effective repletion of the tumor precursor pool with translocated cells; (2) the apparent ability to sustain numerous T(12;15)⁺ clones over extended periods of time; and (3) the continuous generation of subclones of translocated plasma cells.

Materials and methods

Induction and diagnosis of plasmacytomas

Plasmacytomas were induced in plasmacytoma-susceptible, inbred BALB/cAnPt or plasmacytoma-hypersusceptible, congenic BALB/cAnPt.DBA/2N-Idh 1-Pep3 mice by three i.p. injections of 0.5 ml pristane (2,6,10,14-tetramethylpentadecane, Aldrich, Milwaukee, WI, USA) given 60 days apart, as previously described.⁸ One tumor, SIPC 5974, was induced by the i.p. administration of small fragments of a silicone gel.⁴ All mice were bred and maintained at our conventional mouse facility (PerImmune, Rockville, MD, USA) under NCI contract N01-BC-21075. Plasmacytomas were diagnosed by finding 10 or more hyperchromatic, atypical plasma cells in cytofuge specimens of ascitic cells that were obtained by abdominal paracentesis and stained with Wright-Giemsa. Tumor-bearing mice were killed and mesenteries with adjacent tumor nodules and inflammatory granulomata were obtained. Tissue samples were either processed for histologic analysis or snap-frozen in liquid nitrogen for DNA preparation.

Histology and immunostaining

For histology, the granulomatous tissue and the tumor nodules were fixed in a mixture of 70% ethanol/formalin/glacial acetic acid (20/2/1, v/v/v), transferred to 70% ethanol and embedded in paraffin. Four-m-thick sections were stained with hematoxylin and eosin (American Histo Labs, Gaithersburg, MD, USA) to evaluate the distribution of clusters of atypical plasma cells and confirm the diagnosis of plasmacytoma. Additional sections were stained with Giemsa (according to Lennert), PAS (suitable for staining glycoproteins like Ig), and methyl green and pyronine (suitable stain for plasma cells) in order to distinguish normal from atypical plasma cells in granulomas. To visualize reticular fibers in granulomata, silver staining according to Gomori was used. For the immunohistochemical determination of Ig production by plasma cells, tissue sections were stained with a combined avidin-biotin immunoperoxidase technique. Antisera to light-chains and heavy-chain classes that are most relevant for BALB/c plasmacytomas (ie , and ) were used as the primary reagents for immunostaining (Southern Biotechnology Associates, Birmingham, AL, USA). The discrimination between IgA- and IgG-producing plasma cells was a major focus of this study. Therefore, no attempt was made in the majority of mice to distinguish the four -isotypes, 1, 2a, 2b, and 3, and a pooled antiserum for the simultaneous detection of all four of them was used. However, in two cases (4122 and 4128), which were characterized by a very pronounced tissue infiltration with IgG⁺ cells, the -isotypes were distinguished by staining with specific antisera. Sections were counterstained with hematoxylin and lithium chloride and evaluated with an Olympus light microscope (Olympus Optical, Tokyo, Japan).

PCR primers and amplification

Most PCR primers used in this study were designed with the help of the Oligo 4.1 software (National Biosciences, Plymouth, MN, USA). Primer sequences for the detection of recombinations involving the DFL16 and VH7183 gene segments were obtained from the literature.^9,10 Primers with the suffix 'a' were used in one-round PCR amplifications and primers with the suffix 'b' were used as nested primers in the second round of two-round PCR reactions (Table 1). For the preparation of high molecular weight DNA, peritoneal granulomata were processed with a standard phenol/chloroform extraction protocol that included digestion with proteinase K (100 g/ml, Boehringer-Mannheim, Indianapolis, IN, USA) and RNAse A (40 g/ml, Sigma, St Louis, MO, USA). Long-template and high-fidelity PCR methods were employed with the aid of Boehringer kits to detect (1) junctional fragments between Igh and c-myc (ie the molecular indications of T(12;15) translocations); (2) D_HJ_H and V_HD_HJ_H rearrangements; and (3) hybrid isotype switch junctions. PCR amplification was performed on 500 ng aliquots of genomic DNA. Hot-start one-round or two-round nested amplifications were carried out with the following cycling conditions: initial denaturation at 96°C for 5 min with the reaction subsequently being held at 80°C for the addition of polymerase mix (1.75 units per 50 l reaction). This was followed by 25 cycles of template denaturation at 95°C for 15 s, primer annealing at 62°C for 15 s, and primer extension at 68°C for 4 to 7 min in the first cycle with a progressive prolongation of the extension time by 20 s for each of the subsequent cycles. Final extension times amounted to 11.5 to 14.5 min. For the second round of PCR, 1 to 5 l of reaction mixture from round one were used with nested primer pairs for an additional round of 25 cycles. The concentration of dNTPs (200 mM) and Mg²⁺ (1.5 to 1.75 mM final) were the same in both rounds.

DNA sequence analysis of clonal PCR fragments

PCR fragments were gel purified with the assistance of the QIAquick gel extraction kit (Qiagen, Valencia, CA, USA) and sequenced directly using the Femtomole cycle sequencing kit (Promega, Madison, WI, USA). An intermediate cloning step of the PCR fragment was thus avoided. This was particularly important for fragments that contained highly repetitive Ig switch regions, which are notorious for undergoing deletions upon cloning, even when recombination-deficient E. coli host cells are employed. Great care was taken to exclude artifactual PCR results. First, all PCR reactions were performed with appropriate positive controls (genomic DNA obtained from plasmacytoma cell lines that contained known Igh/c-myc junctions) and negative controls (liver and kidney DNA that was devoid of Igh/c-myc junctions). Second, all recombinational fragments from peritoneal granulomata were PCR amplified in quadruplicate and sequenced at least twice to confirm their authenticity. Junctional fragments that could not be reproduced using the same DNA sample were considered artifacts and excluded from further analysis. It is likely that such fragments were derived from individual T(12;15)⁺ cells or microclones of translocated cells that were too small to permit the reproducible detection of clonal Igh/c-myc junction fragments. Third, subfragments were generated from those PCR fragments that were too large (4-10 kbp) to be sequenced easily. The subfragments were produced by 'primer walking', ie repeat PCR amplification of the same tissue sample with a series of staggered PCR primer pairs that annealed progressively closer to the putative Igh/c-myc breaksite, thereby converging on the breaksite from both the 5' end and the 3' end of the chimeric sequence. This approach was empirical, but resulted usually in a reasonably small subfragment (in most cases less than 3 kbp) that could be conveniently sequenced. Thus, most of the junction sites reported here have been re-detected multiple times, which added validity and confidence to the results.

Results

Incidence and diagnosis of plasmacytomas

A total of 17 mice bearing primary peritoneal plasmacytomas were included in this study (Table 2, column 1). The diagnosis of plasma cell tumors was established cytologically and, after the mice were killed, confirmed by histological analysis of stained tissue sections of granulomatous tissue that was obtained from the peritoneal cavity of BALB/c mice previously treated with pristane or silicone gel fragments. The tumor diagnosis was in most cases straightforward and based primarily on the distinct cytological features of the large hyperchromatic plasmacytoma cells that were further characterized by the typical centripetal infiltration pattern observed when the tumor spread from one granuloma to the next. In four randomly selected mice, numbered 3609, 3610, 5974, and 7603, the reliability of the diagnosis of plasmacytoma was tested by the transfer of tumor cell-containing ascites (obtained from the original tumor-bearing animal) into pristane-primed syngeneic recipients. As expected, all four tumors were readily propagated in vivo (Table 2, footnote 'n'). This result indicated that the transferred cells were indeed fully transformed and not confused with premalignant precursor cells of similar morphological appearance.

Staging of plasmacytomas

To estimate the degree of plasmacytoma progression, the amount of tumor cell infiltration in the granulomatous tissue was evaluated histologically. It was found to vary greatly among the 17 mice, ranging from small localized clusters of plasmacytoma cells (defining the 'early stage'), to disseminated tumor infiltrates ('intermediate stage'), to terminal plasma cell tumors that had virtually replaced the entire intraperitoneal granulomatous tissue with a continuous tumor mass ('advanced stage'). Four mice had early stage tumors, six had intermediate stage disease and seven had advanced plasmacytomas (Table 2, column 2). Unlike plasmacytoma progression, the morphology of the tumors showed very limited variation; all plasmacytomas displayed the typical mature plasmacytic phenotype. Occasionally, signs of subtle intra-tumor heterogeneity were revealed upon close inspection of the entire tumor sample. This usually took the form of tumor nodules containing immature plasmablastic cells. In all mice, neoplastic plasma cell disease was accompanied by infiltrates of normal plasma cells, which were seen either as individual cells scattered throughout the granulomata, as clusters, or as foci of plasma cells that included respectively (according to the convention used in our laboratory) less than 10, 10 to 50, or more than 50 cells. Normal plasma cells could be distinguished readily from their neoplastic counterparts with the help of the following parameters: cell size, nuclear morphology, cytoplasmic staining intensity, tissue distribution pattern, and absence of neighboring tissue mast cells (which were seen frequently in the vicinity of neoplastic but not reactive plasma cells).

Detection of T(12;15) translocations by PCR

PCR-amplified chimeric Igh/c-myc junctions, the clonotypic molecular indicators of T(12;15) translocation,^{2,5,6,7,11,12} were used to distinguish among different plasmacytoma clones and their putative precursors. PCR was performed as high-fidelity or long-template amplification with one primer pair for single-round reactions and two nested primer pairs for two-round reactions. Two to 10 kbp long T(12;15)-typical junction fragments were commonly generated. The clonal junction fragments were not only informative concerning the fine structure of the translocation (clonotypic Igh/c-myc breaksite), they were also useful for further characterizing individual B cell clones on the basis of unitary or composite Igh switch regions and D_HJ_H or V_HD_HJ_H rearrangements. Switch regions harbored clonotypic rearrangements that took the form of internal deletions or hybrid switch junctions (see Figure 1, fragments 12.1a-e as examples). Clonotypic D_HJ_H or V_HD_HJ_H recombinations were detected on some Chr T(15;12)-typical products (see fragment 15.1 in Figure 1 and fragments 15.1 and 15.2 in Figure 2). The PCR methods utilized here resulted in the detection of exchanges between c-myc and the following Igh loci: , 2a, 2b, and . Exchanges with and were not observed because dedicated PCR methods have not been developed (/c-myc exchanges are very rare in BALB/c plasmacytomas¹³ and /c-myc exchanges have never been found). Exchanges with 1 and 3 were also not observed in spite of the availability of dedicated PCR methods. The reason for this remained unclear, but one possibility is persistent problems with PCR.

Specificity and sensitivity of PCR

To determine the specificity of the PCR for detecting T(12;15)-typical junction fragments, serial dilution and co-amplification experiments were performed, similar to previously described studies that utilized unique Igh primers,^5,6 switch consensus primers,⁷ or both unique Igh and switch consensus primers in conjunction with hybridization-enriched templates.⁶ To optimize the PCR conditions for the various primer pair combinations listed in Table 1, the following plasmacytomas were chosen as model templates for different types of T(12;15) translocations: TEPC 1194 (C/c-myc), TEPC 1033 (C2a/c-myc), MOPC 21 (C2b/c-myc), and XRPC 24 (C/c-myc).¹⁴ After optimization of PCR, only two, namely the tumor-specific Igh/c-myc junctions of both products of translocation were detected. That is, variant or 'ghost' bands did not occur, which documented the specificity of the PCR. To determine the sensitivity of PCR, tumor DNAs were serially diluted (individually or combinatorially) into liver or kidney DNA. When one-round PCR amplifications of aliquots of 500 ng (corresponding to ~2 ´ 10⁵ cells) of these DNA mixtures were performed in quadruplicate, it was found that approximately 2.5 ng of tumor DNA (corresponding to ~1000 cells) had to be present to reliably generate the expected PCR products in all four reactions. The sensitivity of two-round PCR with nested primer pairs was naturally higher and approached approximately 20 tumor cells among a total of 2 ´ 10⁵ cells. That is, 20 tumor cells were detected in all four reactions when the experiments were performed in quadruplicate. The reproducibility of the PCR deteriorated rapidly at even lower copy numbers of Igh/c-myc rearrangements (<20 tumor cell genomes) and resulted in the failure to reliably detect the expected PCR product in quadruplicates. Another limitation to the sensitivity of PCR, template competition, was noted in co-amplification experiments that utilized the same primer pair to co-detect two or three templates present at identical (low) copy numbers. The co-amplifications resulted frequently in the detection of only one or two PCR products (instead of the expected two or three products), even though all templates could be readily amplified under the same PCR conditions when tested individually. This experience led to the realization that T(12;15)⁺ clones had to be at least moderately expanded (~50-100 tumor cells per 500 ng DNA tissue sample) in order to be reproducibly detectable.

Minor and major clones harboring translocations

T(12;15)-typical fragments that were reproducibly detected after a single round of PCR were postulated to indicate the presence of prevalent or 'major' clones of malignant plasma cells. All mice included in this study were found to contain such clones (see Table 2, column 5 and fragment 15.1 in Figure 1 and fragments 15.2 and 12.1b in Figure 2). T(12;15)-typical fragments that required two-round, nested PCR for detection were postulated to indicate the presence of subordinate or 'minor' clones of malignant plasma cells. The majority of clones described here were clones of that sort. While the distinction between minor and major clones of translocated cells was unquestionably useful for consideration of clonal diversity and clonal expansion, it was also burdened with uncertainties. First, the neoplastic nature of minor clones was not established. While it was reasonable to assume that the high-abundancy Igh/c-myc fragment was the PCR correlate of the histomorphologically prevalent plasmacytoma, it was less clear whether the low-abundancy Igh/c-myc fragments present in the same tissue were predictive of minor plasmacytoma clones or non-malignant clones of plasma/progenitor B cells harboring translocations. Second, the efficacy of the PCR was variable; ie not all Igh/c-myc junctions present at similar copy numbers were detected with equal reliability. The variability appeared to be caused, in addition to the reasons discussed in the previous paragraph, by the sequence context of the amplicon (eg the nature and length of switch regions) and its length. Third, the co-existence of up to eight (see mouse 4132 explained below) distinct recombinational fragments in the same tissue posed another problem, because it was difficult to uncover several junction fragments in the same DNA sample without relaxing the conditions for excluding PCR artifacts. The above-mentioned uncertainties led to the understanding that the distinction between major and minor T(12;15)⁺ clones was limited. Nevertheless, it still appeared meaningful for the purpose of this study, particularly as long as a truly quantitative methods for the accurate enumeration of translocated clones in complex tissues have not been developed.

T(12;15) as a marker of clonal diversity

The results of the evaluation of the T(12;15) translocation in 17 mice bearing primary plasmacytomas are summarized in Table 2, columns 5-8. The determination of the C_H locus that was utilized in the prevalent tumor clone detected by one-round PCR (the 'major' clone) revealed that C was most frequently used in recombinations with c-myc on Chr T(12;15), namely in 13 out of 17 (76.5%) tumors (column 5). C was utilized in four cases (23.5%). Thus, in the sample of primary plasmacytomas studied here, C and C appeared to be the loci of choice for c-myc rearrangements on the c-myc-deregulating product of translocation. The screening for apparently less prevalent or 'minor' clones of translocated cells detected by two-round PCR uncovered the presence of multiple T(12;15)⁺ clones (column 6) and subclones derived from common translocated progenitors (columns 7 and 8). They were found in nine and five mice, respectively. If one considered columns 6-8 together, it was evident that 12 out of 17 (70.6%) plasmacytomas were characterized by clonal diversity. The reason why 'minor' clones were apparently absent in the remaining five mice was not known, but clearly not related to the progression stage of the plasmacytoma (early tumor in mouse 4123; intermediate tumors in mice 4121, 4122 and 4124; advanced tumor in mouse 7609). Possible explanations for the lack of minor clone detection in these mice include the presence of plasmacytomas that contained unusual rearrangements and the limits of the PCR, as alluded to in the preceding paragraph. However, neither of these possibilities has been demonstrated.

Clonal mosaics and intraclonal heterogeneity

Two principal manifestations of clonal diversity were found among malignant plasma cells when the molecular structure of the T(12;15) translocation was used as the clonotypic marker: intraclonal heterogeneity and clonal mosaics. Intraclonal heterogeneity was defined by the presence of translocated subclones derived from common T(12;15)⁺ progenitors. It was effected by either aberrant isotype switching in the vicinity of the c-myc rearrangement ('translocation remodeling') or deletions in translocation breakpoint regions. The clonal multiformity that is indicative of intraclonal diversification is illustrated in Figure 1, using mouse 5974 as an example. Mosaics were defined by the occurrence of multiple translocated clones, which were thought to be independent, ie unrelated to each other with respect to the fine structure of the T(12;15). The clonal multiformity that is indicative of mosaicism is illustrated in Figure 2, using mouse 4132 as an example. Figures 1 and 2 should be inspected in conjunction with Table 1 (for details on PCR primers), Table 2, columns 5-8 (for an overview on clonal diversification), and Table 3 (for DNA sequence analysis of recombination junctions).

Figure 1, panel c depicts schemata of four Chr T(12;15)-typical clonal junction fragments between C and c-myc. They are postulated to indicate descendant subclones of the common C/c-myc⁺ precursor clone, 12.1a (shown in panel b). Fragment 12.1b must have been derived from fragment 12.1a by an aberrant class switch event that joined S and S (indicated by the arrow labeled 'Abnormal isotype switching'). This can be postulated because the clonotypic junction site between E and c-myc was shared in fragments 12.1a and 12.1b. Fragment 12.1c is believed to define a subclone of 12.1b, from which it is distinguished by a small deletion in S (indicated by a small gap and a triangle pointing up) that was found in the vicinity of the S/S joint. The latter was common for fragments 12.1b and 12.1c. Fragment 12.1d is thought to define another subclone of 12.1b; in this case derived by a deletion (depicted by two dashed lines) of the E/c-myc breakpoint region (marked by breaksite 3) and the internal rearrangement in S (marked by breaksite 4 and the grey triangle pointing down). Fragment 12.1e must denote a descendant from fragment 12.1d, which originated by a small secondary deletion in S (indicated by the triangle pointing up). The precursor-and-product relationship between 12.1d and 12.1e was certain because the junction site between S and intron 1 of c-myc (marked by breaksite 5) was shared in both fragments. The findings in mouse 5974 were consistent with the view that BALB/c plasmacytomagenesis is characterized by the clonal diversification of T(12;15)⁺ cells that give rise to progenitor clones by deletions in translocation breakpoint regions (either removing the original breaksite in c-myc or shortening the switch regions in its vicinity) or isotype switch like recombinations (joining S and S in close proximity to c-myc rearrangement).

Figure 2 illustrates the most prominent example of a clonal mosaic that was observed in this study. Mouse 4132 is believed to contain five to seven independent T(12;15)⁺ clones. The exact number of clones remained indeterminate for the reasons stated below. Fragments 12.1a and 12.1b were found to be related in a precursor-and-product relationship, which was mediated by an abnormal class switch recombination that involved three switch regions, S, S2b and S (see arrow pointing down labeled 'Abnormal isotype switching'). None of the six remaining fragments seemed to be related to each other at the fine structural level. They were therefore interpreted to indicate the presence of unique translocated clones. The Chr T(12;15)-typical junction fragment 12.2 could not have been related to clones 12.1a/b because of a different breakpoint in c-myc (cf lines 17 and 20 in Table 3). The five Chr T(15;12)-typical junction fragments must also represent distinct clones for the following reasons: first, fragment 15.2, which could have been derived from fragment 15.1 by the hypothetical internal deletion that is indicated by two dashed lines, was shown to contain a VDJ rearrangement that was distinct from the VDJ recombination observed in fragment 15.1 (cf lines 6-7 and 9-10 in Table 3). This observation effectively excluded a common clonal origin. Second, fragments 15.3 and 15.4 utilized different S regions for recombinations with c-myc, the inverted S2b and S1 regions, respectively. Third, fragment 15.5 originated by a rare c-myc recombination with the native C locus prior to class switching. This interpretation is based on the presence of the small I exon and the 5'-I region, which would have been deleted if isotype switching to C had occurred prior to recombination with c-myc. The considerations offered above argue in favor of seven independent clones co-present in mouse 4132. However, this number could be lower if some of the Chr T(12;15)- and Chr T(15;12)-typical fragments were in fact paired (reciprocal) products of the same translocation event and reside, therefore, in the same clone. Thus, it may be possible that fragments 12.1a and 12.1b correspond to fragments 15.1 or 15.2. Furthermore, fragment 12.2 may be the counterpart to fragment 15.3, since both contained S2b sequences. Because of these uncertainties, the true number of independent clones co-present in mouse 4132 can not be decided; yet, the minimum appears to be five and the maximum, seven.

Ig expression as a marker of clonal diversity

Immunoglobulin production was exploited as an additional marker to assess the clonal diversity of BALB/c plasma cell tumors. Various clones of neoplastic plasma cells residing in the same granulomatous tissue were distinguished on the basis of isotype productions, using immunostaining with monospecific antisera to immunoglobulin heavy or light chains as the tool. Malignant plasma cells were differentiated from their normal counterparts with the assistance of the morphological criteria described in the second paragraph of the Results section. Immunolabeling was found to be particularly valuable for discriminating the prevalent tumor clone against the accompanying subordinate clones of neoplastic plasma cells. The isotype production of the prevalent tumor clone is listed in Table 2, column 3. It shows that the majority of plasmacytomas produced IgA (11/17, 64.7%), five of 17 tumors (29.4%) expressed IgG, and one tumor (1/17, 5.9%) did not produce Ig at all. Of significance, all mice that harbored a predominantly IgA⁺ plasmacytoma were shown to contain additional IgG⁺ clones. Likewise, all mice that harbored a predominantly IgG⁺ plasmacytoma contained additional IgA⁺ clones (Table 2, column 4). These findings led to the conclusion that the majority of primary plasmacytomas are composed of mosaics of IgA- and IgG-producing clones (see Figure 3 for two typical mosaics). Immunostaining for Ig secretion was not helpful, however, for associating isotype production with the molecular data on the T(12;15) translocation. Thus, while it is tempting to speculate that the immunodominant plasmacytomas (eg the IgA⁺ tumors in mice 4132 and 5974) may have contained the most prevalent PCR junction fragments (ie the fragments detected after a single round of PCR; 12.1b in mouse 4132 and 12.1b/c in mouse 5974), it has to be stated that this association has not been established here.

Loss of Ig production and isotype switch variants

The superimposition of serial tissue sections immunostained for IgA, IgG, and IgM indicated that primary BALB/c plasmacytomas may be further diversified by generating rare Ig expression variants. The presence of Ig^- tumor cells in a discrete and clearly demarcated portion of an IgA-secreting plasmacytoma, observed in mouse 4123 (Figure 4a-d), raised the possibility that the Ig^- cells constituted a subclone of the IgA⁺ plasmacytoma. Similarly, the occurrence of IgA⁺ tumor cells in a IgG⁺ tumor nodule, uncovered in mouse 4121 (Figure 4e-j), suggested that the IgA⁺ cells represented isotype switch variants that were derived by untimely class switch recombination from the IgG⁺ plasmacytoma. However, since the clonal relatedness to the parental clone was neither established for the putative Ig^- subclone nor the putative isotype switch variants, additional studies will have to be performed to verify the hypothetical relationship. A number of methods and clonotypic markers, including V_LJ_L/V_HD_HJ_H rearrangements or T(12;15) translocations, could be exploited for this purpose. Allele-specific RT-PCR methods, in which reverse PCR primers specific for different isotypes are paired with tumor specific V_H primers, have been used successfully in multiple myeloma research to demonstrate, by virtue of a shared clonotypic V_HD_HJ_H rearrangement, that myeloma clones can give rise to isotype switch variants. Allele-specific RT-PCR may also offer itself as the method of choice for BALB/c plasmacytomas; indeed, its combination with a precise microsampling tool (eg laser capture microdissection) may be effective for analyzing putative in situ Ig expression variants like the ones described above.

Discussion

Igh/c-myc junction fragments amplified by PCR are not only accurate indicators of reciprocal T(12;15) chromosomal translocations,² they also offer arguably the most useful molecular parameter currently available for evaluating clonal diversity in the course of BALB/c plasmacytomagenesis. The reasons for this usefulness include, first, the reality that the T(12;15) translocation is the predominant type of c-myc-activating rearrangements in mouse plasma cell tumors, occurring with an incidence of approximately 80%.¹⁴ Second, the translocation is thought to be specific for the B cell lineage. The presumed lineage specificity is important because it allows the unambiguous association between the translocation events detected in complex tissues and a particular cell type, namely B cells and their plasma cell progenitors. Third, the fine structure of the T(12;15) translocation is believed to be unique and clonotypic because no Igh/c-myc junction has thus far been found to be shared between two plasmacytomas. This distinctiveness can be utilized effectively as a molecular fingerprint for monitoring the tissue distribution and migration of translocated clones.⁷ Finally, recombinational fragments between Igh and c-myc can be detected by PCR with remarkable sensitivity.^{2,5,6,7,11,12} This makes possible the in vivo detection of small translocated clones that in all likelihood would have been missed by alternative detection methods, such as Southern blotting, FISH, or conventional cytogenetics. An additional benefit of PCR beyond the detection of translocation events per se is the possibility to specify the Igh locus utilized for a particular c-myc rearrangement. The differentiation between C and more distal C_H loci has proved to be particularly instructive because it may provide a clue for identifying the still elusive plasmacytoma precursor cell.¹¹

Using the fine structure of the T(12;15) translocation as the molecular marker to define clonality, two distinct forms of clonal diversity were uncovered among plasmacytoma cells and their precursors: multiple translocated clones that were unrelated to each other (mosaics) and subclones that were generated from common translocated progenitors (intraclonal diversification). First, in nine out of 17 (53%) mice bearing primary plasmacytomas, multiple independent T(12;15)⁺ clones were detected. The clones were scored as independent when the DNA sequence analysis of the breakpoint regions of both products of translocation failed to reveal an apparent precursor-and-product relationship. A possible limitation of this definition is caused by the chance that some putatively independent clones (as defined by their unique Igh/c-myc junction sequences) might nevertheless have been derived from each other, but their relatedness could not be established with the relatively small translocation breakpoint region as the exclusive marker. This situation may have resulted in an overestimation of the frequency of independent clones reported here. On the other hand, the PCR methods used in this study were restricted to detecting those Chr T(12;15)-typical rearrangements of c-myc that are common for established plasmacytomas: ie with C, C, C2a, and C2b.¹⁴ This restriction may have resulted in the underestimation of the true number of independent clones, since all presumptive clones that used C, C1, C3 or C in their recombinations with c-myc were left undetected. C1/c-myc and C3/c-myc rearrangements may have been missed because of PCR problems (see Results section, third paragraph). C/c-myc and C/c-myc rearrangements are very rare in BALB/c plasmacytomas; nevertheless, their presence in the sample studied here can not be excluded. Thus, while the true fraction of BALB/c mice that harbored multiple clones and the true incidence of these clones in individual mice cannot be defined accurately at present, the available results suggest that these numbers must be substantial.

In five out of 17 (29.5%) mice, subclones thought to be derived from common T(12;15)⁺ progenitors were found. Deletional mutagenesis in translocation breakpoint regions was identified as the main mechanism by which such subclones were generated. Curiously, the deletions seemed to be confined to Chr T(12;15), the functionally important product of translocation that harbored the deregulated c-myc gene. It is currently unclear whether the apparent restriction to one product of translocation reflects the reality of the phenomenon or is an artifact of preferentially detecting deletions on one allele. We favor the first explanation and postulate that translocation breakpoint regions may be less stable on Chr T(12;15) than on Chr T(15;12), the reciprocal and presumably inconsequential product of translocation. The reason for the putative allele-specific instability is not known, but it is possible that it is related to the known instability of the Igh locus in plasmacytoma cells.^15,16,17,18 Another attempt to explain the instability considers a role for the 3'-C enhancers, which are always allocated to Chr T(12;15) by the translocation event. Since the 3'-C enhancers have been invoked before as possible control elements for isotype switching,¹⁹ illegitimate recombinations in the C_H gene cluster^20,21,22 and, indeed, for the general activation of the chromatin domain containing the Igh locus,²³ it seems plausible that the enhancer may also facilitate the deletional mutagenesis in Chr T(12;15)-typical Igh/c-myc junctions. This mutagenesis, which appeared to take three distinct forms, will be discussed in the following three paragraphs.

The first type of deletion seemed to 'chip away' at the unitary or composite switch regions that were frequently present in the vicinity of the Igh/c-myc breaksites; however, it left the breakpoints themselves unchanged (see fragments 12.1c and 12.1e in Figure 1c). This kind of deletion showed a striking similarity to deletions in unitary S and hybrid S/S regions that have been described to occur in normal B cells prior to primary and sequential switch recombinations, respectively.²⁴ Switch region deletions of this sort are believed to lock normal B cells into producing a particular Ig heavy chain, thereby accomplishing a state of 'isotype stabilization'.²⁴ The B cell with the stabilized isotype is, according to this concept, unable to undergo primary switch recombination (if S is deleted) or secondary/sequential switch recombination (if a combined switch region, such as S/S, is deleted) because the donor substrate for switching has been significantly reduced or completely lost. Our data suggest that isotype stabilization may have an analog in transformed plasma cells, in which translocation breakpoint regions may be rendered more stable after the unitary or composite switch regions residing on Chr T(12;15) have been reduced in size or eliminated. The functional importance of this phenomenon, which could be coined 'translocation stabilization', is presently unknown. However, it is conceivable that translocated B cells with a 'stabilized' c-myc rearrangement may have a selective advantage because they are less likely to undergo potentially deleterious secondary rearrangements of the recombinogenic (!) switch regions in the vicinity of the Igh/c-myc junction. Such secondary rearrangements, which may be manifested as deletions, inversions or three-way recombinations, may reach into the protein-encoding portion of c-myc or otherwise interfere with the constitutive transcription of the c-myc gene, the key requirement for the immortalization/proliferation of the translocated clone.

The second form of deletion that was observed on Chr T(12;15) appeared to target the Igh/c-myc breaksite with its adjoining flanks. This type of deletion can create a dilemma for the student of the clonal evolution of translocated cells, because it obstructs the association of newly generated subclones with their precursors by virtue of removing the original Igh/c-myc junction and producing a secondary, shortened Igh/c-myc junction. This situation is exemplified by the fragment 12.1d/e in Figure 1. In general, deleted subclones of this type can only be related to their putative precursors when, fortuitously, unique mutations (eg base substitution mutations) or distinctive rearrangements (eg hybrid switch junctions) are preserved in the breakpoint flanks adjacent to the deleted stretch of DNA. Such mutations and rearrangements have been documented in previous work.⁵

The third type of deletion removed the heavy-chain gene cluster almost entirely, effecting a loss of approximately 180 kb of DNA. The magnitude of the loss clearly distinguished this deletion from the two above-described deletions, which were much more limited in scope by removing no more than approximately 3 kb of genomic sequences. The 180-kb macrodeletions were observed in clones in which an initial exchange of c-myc with C was converted into a secondary exchange of c-myc with C. In this situation, all C_H loci residing between C, the most upstream CH locus, and C, the most distal C_H locus, were eliminated from the Chr 12-derived flank of the translocation breakpoint region. Macrodeletions, which left the original break in c-myc as unchanged as the above-discussed microdeletions in switch regions, were observed in a significant fraction of the plasmacytoma sample, in five out of 17 tumors (29%, see the 12.1b fragments in Figures 1 and 2 as examples).

The molecular mechanisms for the above-noted deletions are not currently known. The macrodeletions, which were originally described as 'deletional remodeling of T(12;15) translocation breakpoint regions',¹¹ have been ascribed to aberrant isotype switching, but in reality it remains unclear whether the deletions were indeed performed by the putative switch recombinase (just like in normal class switching) or simply by general DNA double-strand break repair. The latter could effect illegitimate recombination between switch regions by non-homologous end joining or similar mechanisms, and produce thereby a chimeric switch region that could not be distinguished at the fine structural level from one generated by isotype switching.²⁵ It is also conceivable that the two forms of microdeletions: ie removal of switch regions and elimination of original breaksites in c-myc, were caused by double-strand break repair. Alternatively, their origin may be related to the peculiar deficiency of BALB/c mice to perform transcription-coupled DNA repair. Since this deficiency appears to affect loci that are prominently involved in T(12;15) translocations (eg Igh and c-myc),^26,27 it is tempting to speculate that BALB/c mice may possess a dysfunctional gene-specific repair complex that is somehow responsible for the predisposition to undergo deletions in translocation breakpoint regions. However, this has not been demonstrated.

The findings on the coexistence of IgA⁺ and IgG⁺ plasmacytoma clones in all tumor-bearing mice were not only consistent with the concept of ongoing clonal diversification in the course of plasmacytomagenesis, they also confirmed and extended results from previous studies in which Ig expression was used as the marker for clonal diversity. In one experiment, focal proliferations of atypical, premalignant plasma cells (plasmacytic foci) were studied immunohistochemically with isotype-specific sera in 62 BALB/c mice that had been treated with pristane.²⁸ Fifteen mice (24.2%) were observed to be biclonal, three (4.8%) triclonal and two (3.2%) tetraclonal with respect to Ig heavy chain expression. Significantly, in many cases there was a preponderance of one clonotype (numerous foci producing IgA), but only a single representative of a second clonotype (one focus expressing IgG). This situation was interpreted to reflect the emergence of a dominant tumor clone at a stage of oncogenesis at which one or more minor clones were still present.²⁸ In another experiment that included as many as 576 mice bearing primary plasmacytomas, the monoclonal immunoglobulins (paraproteins) in the ascitic fluid were determined by agarose- and immunoelectrophoresis. In spite of the relative insensitivity of these methods, 54 (9.4%) mice were found to contain two or more distinct paraproteins.²⁹ This result provided an indication that a substantial fraction of primary plasmacytomas must consist of several distinct paraprotein-producing clones. In a third early experiment, multiple pristane-induced granulomata containing putative tumor nodules were obtained from one original donor mouse and transferred individually into several syngeneic recipients. In three cases, distinct tumors producing different paraproteins were established from the same donor. In the most dramatic case, as many as five independent cell lines were derived from one mouse.³⁰ This result was particularly instructive because it went beyond simply indicating that BALB/c mice may harbor mosaics of T(12;15) translocated plasma cells. Instead, it strongly suggested that a significant number of plasmacytic clones, which are diagnosed by the histomorphologist as aberrant, have already completed malignant transformation, and are thus able to give rise to plasmacytomas upon transplantation.

The present study has demonstrated that primary BALB/c plasmacytomas are usually characterized by an ongoing clonal diversification that is manifested by the presence of multiple T(12;15)⁺ tumor clones (mosaics) and the generation of subclones from common translocated progenitors (intraclonal heterogeneity). Plasmacytomas can therefore be added to the plethora of mouse models of human cancer (B cell lymphomas,³¹ T cell lymphomas,³² and other malignancies³³) that have been utililized successfully to study clonal diversity during oncogenesis. However, the possibility to associate clonal diversification processes with a pool of tumor/tumor progenitor cells, which is clearly defined by clonotypic T(12;15) translocations and monoclonal Ig production, may be a unique advantage of the BALB/c plasmacytoma system. This advantage should be exploited in future research to attain two important goals. First, the ability to undergo efficient clonal diversification may constitute a critical determinant of the genetic susceptibility to plasmacytomagenesis, a very unusual trait for most inbred strains of mice but a characteristic phenotype for BALB/c mice.³⁴ Studies on the genetic control of oligoclonality may thus lead to the identification of novel plasmacytoma susceptibility alleles that may have broad implications for plasma cell neoplasias in general, including human diseases. Second, the results on clonal diversification at the terminal stage of plasmacytoma development presented here - viewed in conjunction with previous findings on clonal diversity at early stages of plasmacytoma development^5,6,7 - raise the intriguing possibility that the deregulated c-myc gene may function as the driving force of clonal diversity: eg by inducing genomic instability in plasmacytoma precursors. A role for c-myc as a mutator gene in B-lineage and plasma cells has yet to be demonstrated, but studies in other cell lineages strongly support such a hypothesis.^35,36,37,38

Acknowledgements

We thank Dr Michael Potter for supporting this work, sharing his expertise on the histogenesis of plasmacytoma development and making many additional helpful suggestions in the course of the experiments. We gratefully acknowledge the contributions of Wendy duBois to animal husbandry. We are grateful to Drs Lynne D Rockwood, Rebecca Liddell and J Frederic Mushinski for reading the manuscript and providing editorial assistance.

1 Shen-Ong GL, Keath EJ, Piccoli SP, Cole MD. Novel myc oncogene RNA from abortive immunoglobulin-gene recombination in mouse plasmacytomas. Cell 1982; 31: 443-452, MEDLINE

2 Janz S, Muller J, Shaughnessy J, Potter M. Detection of recombinations between c-myc and immunoglobulin switch alpha in murine plasma cell tumors and preneoplastic lesions by polymerase chain reaction. Proc Natl Acad Sci USA 1993; 90: 7361-7365, MEDLINE

3 Potter M, MacCardle RC. Histology of developing plasma cell neoplasia induced by mineral oil in BALB/c mice. J Natl Cancer Inst 1964; 33: 497-515,

4 Potter M, Morrison S, Wiener F, Zhang XK, Miller FW. Induction of plasmacytomas with silicone gel in genetically susceptible strains of mice. J Natl Cancer Inst 1994; 86: 1058-1065, MEDLINE

5 Müller JR, Potter M, Janz S. Differences in the molecular structure of c-myc-activating recombinations in murine plasmacytomas and precursor cells. Proc Natl Acad Sci USA 1994; 91: 12066-12070, MEDLINE

6 Müller JR, Jones GM, Potter M, Janz S. Detection of immunoglobulin/c-myc recombinations in mice that are resistant to plasmacytoma induction. Cancer Res 1996; 56: 419-423, MEDLINE

7 Müller JR, Jones GM, Janz S, Potter M. Migration of cells with immunoglobulin/c-myc recombinations in lymphoid tissues of mice. Blood 1997; 89: 291-296, MEDLINE

8 Potter M, Wax JS. Peritoneal plasmacytomagenesis in mice: comparison of different pristane dose regimens. J Natl Cancer Inst 1983; 71: 391-395, MEDLINE

9 Rolink A, Haasner D, Nishikawa SI, Melchers F. Changes in frequencies of clonable preB cells during life in different lymphoid organs of mice. Blood 1993; 81: 2290-2300, MEDLINE

10 Ehrlich A, Martin V, Müller W, Rajewsky K. Analysis of the B-cell progenitor compartment at the level of single cells. Curr Biol 1994; 4: 573-583, MEDLINE

11 Kovalchuk AL, Müller JR, Janz S. Deletional remodeling of c-myc-deregulating translocations. Oncogene 1997; 15: 2369-2377, MEDLINE

12 Müller JR, Janz S, Potter M. Differences between Burkitt's lymphomas and mouse plasmacytomas in the immunoglobulin heavy chain/c-myc recombinations that occur in their chromosomal translocations. Cancer Res 1995; 55: 5012-5018, MEDLINE

13 Owens JD Jr, Finkelman FD, Mountz JD, Mushinski JF. Nonhomologous recombination at sites within the mouse JH-C delta locus accompanies C mu deletion and switch to immunoglobulin D secretion. Mol Cell Biol 1991; 11: 5660-5670, MEDLINE

14 Potter M, Wiener F. Plasmacytomagenesis in mice: model of neoplastic development dependent upon chromosomal translocations. Carcinogenesis 1992; 13: 1681-1697, MEDLINE

15 Baar J, Shulman MJ. The Ig heavy chain switch region is a hotspot for insertion of transfected DNA. J Immunol 1995; 155: 1911-1920, MEDLINE

16 Oancea AE, Tsui FW, Shulman MJ. Targeted removal of the mu switch region from mouse hybridoma cells. A test of its role in gene expression in the endogenous IgH locus. J Immunol 1995; 155: 5678-5683, MEDLINE

17 Baar J, Pennell NM, Shulman MJ. Analysis of a hot spot for DNA insertion suggests a mechanism for Ig switch recombination. J Immunol 1996; 157: 3430-3435, MEDLINE

18 Buzina A, Shulman MJ. An element in the endogenous IgH locus stimulates gene targeting in hybridoma cells. Nucleic Acids Res 1996; 24: 1525-1530, MEDLINE

19 Cogne M, Lansford R, Bottaro A, Zhang J, Gorman J, Young F, Cheng HL, Alt FW. A class switch control region at the 3' end of the immunoglobulin heavy chain locus. Cell 1994; 77: 737-747, MEDLINE

20 Webb CF, Cooper MD, Burrows PD, Griffin JA. Immunoglobulin gene rearrangements and deletions in human Epstein-Barr virus-transformed cell lines producing different IgG and IgA subclasses. Proc Natl Acad Sci USA 1985; 82: 5495-5499, MEDLINE

21 Lewis SM, Agard E, Suh S, Czyzyk L. Cryptic signals and the fidelity of V(D)J joining. Mol Cell Biol 1997; 17: 3125-3136, MEDLINE

22 Bailey SN, Rosenberg N. Assessing the pathogenic potential of the V(D)J recombinase by interlocus immunoglobulin light-chain gene rearrangement. Mol Cell Biol 1997; 17: 887-894, MEDLINE

23 Arulampalam V, Eckhardt L, Pettersson S. The enhancer shift: a model to explain the developmental control of IgH gene expression in B-lineage cells. Immunol Today 1997; 18: 549-554, Article MEDLINE

24 Zhang K, Cheah HK, Saxon A. Secondary deletional recombination of rearranged switch region in Ig isotype-switched B cells. A mechanism for isotype stabilization. J Immunol 1995; 154: 2237-2247, MEDLINE

25 Radbruch A, Burger C, Klein S, Muller W. Control of immunoglobulin class switch recombination. Immunol Rev 1986; 89: 69-83, MEDLINE

26 Beecham EJ, Jones GM, Link C, Huppi K, Potter M, Mushinski JF, Bohr VA. DNA repair defects associated with chromosomal translocation breaksite regions. Mol Cell Biol 1994; 14: 1204-1212, MEDLINE

27 Beecham EJ, Owens JD, Shaughnessy JD Jr, Huppi K, Bohr VA, Mushinski JF. Decoupling of DNA excision repair and RNA transcription in translocation breaksite regions of plasmacytoma-susceptible BALB/cAnPt mice. Carcinogenesis 1997; 18: 687-694, MEDLINE

28 Potter M, Mushinski EB, Wax JS, Hartley J, Mock BA. Identification of two genes on chromosome 4 that determine resistance to plasmacytoma induction in mice. Cancer Res 1994; 54: 969-975, MEDLINE

29 Morse HC, Pumphrey JG, Potter M, Asofsky R. Murine plasma cells secreting more than one class of immunoglobulin heavy chain. I. Frequency of two or more M-components in ascitic fluids from 788 primary plasmacytomas. J Immunol 1976; 117: 541-547, MEDLINE

30 Potter M. Plasma cell neoplasia in a single host: a mosaic of different protein-producing cell types. J Exp Med 1962; 115: 339-356,

31 Tang JC, Ho FC, Chan AC, Srivastava G. Clonality of lymphomas at multiple sites in SJL mice. Lab Invest 1998; 78: 205-212, MEDLINE

32 Newcomb EW. Clonal evolution of N-methylnitrosourea-induced C57BL/6J thymic lymphomas by analysis of multiple genetic alterations. Leukemia Res 1997; 21: 189-198,

33 Habano W, Sugai T, Nakamura S. Mismatch repair deficiency leads to a unique mode of colorectal tumorigenesis characterized by intratumoral heterogeneity. Oncogene 1998; 16: 1259-1265, MEDLINE

34 Potter M, Pumphrey JG, Bailey DW. Genetics of susceptibility to plasmacytoma induction. I. BALB/cAnN (C), C57BL/6N (B6), C57BL/Ka (BK), (C times B6)F1, (C times BK)F1, and C times B recombinant-inbred strains. J Natl Cancer Inst 1975; 54: 1413-1417, MEDLINE

35 Mai S, Fluri M, Siwarski D, Huppi K. Genomic instability in MycER-activated Rat1A-MycER cells. Chromos Res 1996; 4: 365-371,

36 Mai S, Hanley-Hyde J, Fluri M. c-Myc overexpression associated DHFR gene amplification in hamster, rat, mouse and human cell lines. Oncogene 1996; 12: 277-288, MEDLINE

37 Davis CD, Dacquel EJ, Schut HA, Thorgeirsson SS, Snyderwine EG. In vivo mutagenicity and DNA adduct levels of heterocyclic amines in Muta mice and c-myc/lacZ double transgenic mice. Mutat Res 1996; 356: 287-296, MEDLINE

38 Felsher DW, Bishop JM. Transient excess of MYC activity can elicit genomic instability and tumorigenesis. Proc Natl Acad Sci USA 1999; 96: 3940-3944, Article MEDLINE

Figure 1  'Translocation remodeling' by abnormal isotype switching in a mouse (5974) that bore an advanced primary plasmacytoma. (a) Maps of the unrearranged c-myc gene on Chr 15 and the part of the Igh gene complex on Chr 12 that spans the JH region and the first exon of C. The three exons of c-myc are designated by open boxes that are numbered. The intronic Igh enhancer, the 3.5-kbp switch region and exon 1 of C are illustrated by boxes labeled E, S, and C1, respectively. The JH region, which contains four functional gene segments (JH1-4, vertical black lines) and one pseudogene (JH, vertical grey line), is marked JH. Centromeres are denoted by filled circles at the right end of chromosomal maps. Chrs 12 and 15 are aligned at the postulated cross-over point of the translocation, which is indicated by a vertical dashed line that is labeled 'T(12;15)'. (b) Structural schemas of the observed recombinational fragments between E and the first exon of c-myc after the reciprocal exchange. The translocation produced two chimeric chromosomes, Chr T(12;15), which is known to be critical for plasmacytomagenesis because it harbors the transcriptionally deregulated, protein-encoding portion of c-myc (exons 2 and 3), and Chr T(15;12), which is believed to be irrelevant for oncogenesis. The genetic exchange was found to be nearly precise at the DNA sequence level because it resulted in a loss of just 20 bp and 35 bp in c-myc and E, respectively. The annealing sites (denoted by arrowheads) and designations of PCR primers (see Table 1 for sequence information) used to detect the chimeric PCR fragments 15.1 and 12.1 a are shown. The PCR fragments, which are indicated by grey horizontal bars above and below the maps of Chrs T(15;12) and T(12;15) respectively, span several kb of the translocation breakpoint region and include four recombination breaksites, which are depicted by numbered stars: the translocation breakpoints (2 and 3), the DJ recombination on Chr T(15;12) (1), and the breaksite of an approximately 500 bp long internal deletion in S (4), which is also depicted by a triangle pointing down. (c) Schemata of four additional Chr T(12;15)-typical junction fragments, in which c-myc was found to be recombined with the most 3' locus of the Igh gene cluster, C. The four fragments, designated 12.1b-12.1e, are postulated to be secondary recombinants derived from the C/c-myc precursor fragment, 12.1a, by an aberrant class switch event that joined S and S (see Results section for further details).

Figure 2  Multiple independent T(12;15)⁺ clones (mosaic) in a mouse (4132) that harbored an early primary plasmacytoma. Shown at the top are molecular schemas of five Chr T(15;12)-typical Igh/c-myc junction fragments, 15.1-15.5, and depicted at the bottom are three Chr T(12;15)-typical fragments, 12.1a-12.2. The PCR indicator fragments that were used for DNA sequencing are depicted by horizontal bars above and below the maps of Chrs T(15;12) and T(12;15), respectively. The intronic Igh enhancer, switch regions participating in genetic exchanges with c-myc, the small exon for sterile transcripts of C, and the first exon of the C2b gene are illustrated by boxes labeled E, S2b, S, S, I, and C2b1, respectively. Centromeres are denoted by filled circles at the right end of chormosomal maps. The three exons of c-myc are indicated by open boxes that are labeled. All fragments are aligned along the T(12;15) translocation breakpoint, which is indicated by a vertical dashed line. See the Results section for an explanation why mouse 4132 is believed to harbor five to eight translocated clones.

Figure 3  Mosaics of independent clones of plasma cells in primary plasmacytomas, as defined by isotype production. Immunostained tissue sections of the granulomatous tissue obtained from mouse 4132 are shown in a-f, as an example of a mosaic of neoplastic plasma cells that consisted of a prevalent, IgA-producing plasmacytoma and several subordinate, IgG-expressing clones. An example of the opposite, a mouse (4130) that harbored a dominant, IgG-producing plasmacytoma in conjunction with smaller, IgA-expressing clones of apparent tumor cells, is shown in panels g-l. (a) Pristane-induced, inflammatory granuloma that is heavily infiltrated by a primary, IgA/⁺ plasmacytoma that is labeled brown (, 10 ´); (b) enlargement of the area indicated by a rectangle in panel a (IgA, 20 ´); (c) subordinate, IgG2b⁺ clone (IgG2b, 20 ´); (d) enlargement of the area depicted by a rectangle in panel c, showing IgG2b-expressing tumor cells including one cell that is undergoing mitosis (arrow, IgG2b, 20 ´); (e) subordinate, IgG2a⁺ clone that appears to form two columns of infiltrating tumor cells (arrows, IgG2a, 20 ´); (f) small clone of morphologically atypical, IgG1-producing plasma cells, whose nature as either fully transformed tumor cell or tumor precursor cells can not be determined unambiguously by histological analysis (IgG1, 20 ´); (g) overview of a large portion of inflammatory granulomata that are infiltrated by a primary, pristane-induced, IgG⁺ plasmacytoma (IgG, 2 ´); (h) the enlargement of the area indicated by a rectangle in panel g demonstrates that the portion of the tumor nodule shown to the left of the dashed line is positive for IgG, while the portion to the right of the line is not (IgG, 10 ´); (i) the further enlargement of the image shown in panel h depicts the sharp demarcation line between the IgG⁺ and the IgG^- portion of the tumor infiltrate (IgG, 20 ´); (j) overview of the same granuloma tissue shown in panel g, illustrating the presence of a small, IgA⁺ plasmacytoma (box) in the vicinity of a much larger IgG-expressing tumor (2 ´, IgA) (k) the enlargement of the area indicated by a rectangle in panel j demonstrates that the IgG^- plasmacytoma cells that can be seen in the right-hand area of panel h are clearly positive for IgA (IgA, 10 ´); (l) a further enlargement of the image shown in panel k depicts the marked transition between the IgG⁺ and the IgA⁺ portion of the tumor nodule (IgA, 20 ´).

Figure 4 Intraclonal heterogeneity in primary plasmacytomas, as suggested by the apparent loss of Ig production (panels a to d) and the occurrence of putative isotype switch variants (panels e to j). See last paragraph of the Results section why the presence of intraclonal heterogeneity was suggested by the findings illustrated here, but not proven. Panels a to d illustrate the admixture of IgA-secreting tumor cells with deviants that may have lost the ability to produce Ig. All tissue sections were immunostained for IgA. (a) The overview of the granulomatous tissue (OG) that developed in the mesentery (Mes) adjacent to the small gut in mouse 4123 is presented. A medium-sized plasmacytoma nodule (PCT) which borders on the lamina propria of the gut (LP) and contains a cluster of strongly IgA⁺ neoplastic plasma cells is indicated by a rectangle (2 ´). Smaller aggregates of IgA⁺ tumor cells that are scattered in the granulomata (arrows), as well as numerous normal IgA-producing plasma cells that reside in the lamina propria of the villi (arrowhead), are co-present. Panels b, c and d display progressive enlargements (10 ´, 20 ´ and 40 ´, respectively) of the plasmacytoma nodule that has been marked by the rectangle in panel a. The intimate association of IgA⁺ and IgA^- tumor cells can be seen. The IgA^- cells were also found to be negative for IgG and IgM (results not shown). IgD and IgE production were not determined, since the production of these isotypes by BALB/c plasmacytomas is extremely unusual. Panels e to j demonstrate the occurrence of putative isotype switch variants. The overview of a pristane-induced, inflammatory granuloma that has been completely replaced by an IgG-secreting plasmacytoma that developed in mouse 4121 is shown in panels e (IgG), f (IgA) and g (IgM) at the same magnification (10 ´). (h) The enlargement of the area indicated by a rectangle in panel a, demonstrates that the tumor cells produced IgG (40 ´). (i) The enlargement of the area marked by a rectangle in panel f, reveals the existence of six IgA⁺ tumor cells in the center of the plasmacytoma nodule, a pair and a quartet of IgA-stained cells that have been labeled brown and marked by arrowheads (IgA, 20 ´). (j) The same pair of IgA-producing plasmacytoma cells, already depicted in panel i, is shown in a higher power view (IgA, 40 ´).

Table 1  Sequence and position of PCR primers

Table 2  Clonal diversity in primary BALB/c plasmacytomas

Table 3  Flanking sequences around recombinational breakpoints illustrated in Figures 1 and 2